The recent use of vasodilators to improve ventricular function in acute myocardial infarction led us to investigate the effects of nitroglycerin, nitroprusside, and phentolamine on coronary collateral flow. Dogs were studied 2-4 wk after an ameroid constrictor was placed around the left anterior descending (LAD) coronary artery. Retrograde flow and peripheral coronary pressure were measured from a cannula inserted in the LAD distal to the ameroid. Systemic arterial pressure was held constant by an aortic cuff. When administered intracoronary (i.c.), nitroglycerin, 0.3-100 mug/min, or nitroprusside, 3-100 mug/min, produced quantitatively similar, dose-dependent increases in retrograde flow. Neither drug, i.c., changed peripheral coronary pressure. Nitroglycerin, 3-300 mug/min, intravenous (i.v.), produced dose-dependent increases in retrograde flow; nitroprusside, i.v., increased retrograde flow only in high doses (100-300 mug/min). Nitroglycerin and nitroprusside, i.v., produced similar increases in peripheral coronary pressure. Phentolamine, 1-300 mug/min, i.v., decreased retrograde flow, and did not change peripheral coronary pressure. Nitroprusside was considerably more potent than nitroglycerin in decreasing systemic arterial pressure and in reducing total coronary resistance. Thus, (a) although i.c. nitroglycerin and nitroprusside produce similar effects on collateral function, i.v. nitroglycerin is more effective than i.v. nitroprusside in augmenting collateral flow; (b) phentolamine has deleterious effects on collateral function; and (c) the relative vasodilator potencies of nitroglycerin and nitroprusside vary in different vascular beds; thus, for a given reduction in systemic arterial pressure, nitroprusside is less effective in increasing retrograde flow.